Integrate `churn` Gem into Rubycritic to Speed Up RubyCritic Churn Computation

[hemalvarambhia]

Problem: I use rubycritic on some repositories and some take 15+ minutes, which seems similar to #490.

Solution: I @Greg-Myers-SB found this gem by Dan Mayer (see Churn's Github page) and found it was a lot faster. Can using the said gem happen?

[Greg-Myers-SB]

On the same codebase.

Rubycritic's Churn:

Churn gem:

Churn gem is significantly faster, we might be able to make an adapter for other Churn sources and figure out what the major differences in results are.

[etagwerker]

Solution: I found this gem by Dan Mayer (see Churn's Github page) and found it was a lot faster. Can using the said gem happen?

Yes, I'm all for that. As a co-maintainer, if we can integrate with churn with minimal impact to current users I'm all for it. Thoughts, @nunosilva800?

Churn gem is significantly faster, we might be able to make an adapter for other Churn sources and figure out what the major differences in results are.

I would be inclined to release v5.x for a change like this one. That way we can flag it to rubycritic users that there is a big change getting released.

Ideally we want to keep the rubycritic test suite as is when switching to churn -- I'm not totally sure that there are good enough tests for the churn behavior though.

[Greg-Myers-SB]

Yep, of course it's not just speed, the churn gem is not a 1-1 with the current implementation and we may spend some time into looking at what the differences are.
Edit: I saw on the linked issue that support for HG/Mercurial/SVN are a topic.

[hemalvarambhia]

@Greg-Myers-SB I saw that it has a hg analyser module. I thought Hg and mercurial are the same. How wrong am I?

[hemalvarambhia]

Ideally we want to keep the rubycritic test suite as is when switching to churn -- I'm not totally sure that there are good enough tests for the churn behavior though.

@etagwerker and what kind of good enough is that? Is this something you would like addressed?

Edit: I'd like to credit @Greg-Myers-SB for finding the gem.

[etagwerker]

@etagwerker and what kind of good enough is that? Is this something you would like addressed?

I'm interested in seeing if it considerably changes the churn vs complexity graph in the projects you're working with. Once you have a branch that uses churn, I can test with my own applications and then we can discuss next steps then.

[etagwerker]

Also, just FYI, I don't think we need a release of churn to see this working: danmayer/churn#88 (comment)

I envision the integration will not use churn's JSON output, but it will integrate programmatically with churn

[hemalvarambhia]

Also, just FYI, I don't think we need a release of churn to see this working: danmayer/churn#88 (comment)

I envision the integration will not use churn's JSON output, but it will integrate programmatically with churn

I agree. That would be a huge time saver.

[hemalvarambhia]

I have a small PR (#499) , in which I add a learning test to understand how the gem works. In case anyone's not familiar with the idea of a learning test, it is to way to gain clarification on how an external library and has the bonus of documenting assumptions that can be tested when the gem is updated.

[hemalvarambhia]

@etagwerker #499 contains a learning test (Test-Driven Development By Example, Kent Beck) I thought I could add as a safe microstep towards contributing to Rubycritic. It is small enough to throw away if I need to start over and hopefully easy for you to review and safe enough consider making part of the codebase.

[mateusdeap]

My PR doesn't really integrate churn, but with some help from @hmdros and @julioalucero, we found a simple change that makes churn run significantly faster. The PR has some screen shots.

[Greg-Myers-SB]

Wow that new churn implementation is easily 10x faster on the codebase in my original screenshot. 👏 Well done!